Purchase this article with an account.
Miaomiao Yu, Alex Wade; Evidence for non-linear interactions between endogenous cortical rhythms and periodic chromatic inputs. Journal of Vision 2019;19(8):11. doi: https://doi.org/10.1167/19.8.11.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Rapid, periodic visual stimuli generate time-locked, ‘frequency tagged’ steady state visually evoked potentials (SSVEPs) at the scalp. The amplitudes of the EEG responses at harmonics of the input frequency provide a rich source of information about the neuronal computations that are driven by these ‘exogenous’ stimuli.
The human brain also generates endogenous rhythms: spontaneous, ongoing oscillations of neuronal population firing that have little phase synchrony. Such oscillations are typically invisible in an SSVEP analysis because their phase is random from moment to moment and so their time-locked amplitude averages to zero.
Here we ask whether we can detect the interaction of periodic chromatic visual inputs with ongoing endogenous rhythms. Specifically, we ask whether we can detect systematic changes in response amplitude or phase at frequencies other than those driven directly by a single periodic input. Such changes would be indications of non-linear interactions (‘intermodulation’) between exogenous and endogenous rhythms. Because some researchers have identified slow endogenous rhythms confined to particular pre-cortical visual pathways, we also ask whether these putative non-harmonic responses change as a function of stimulus color.
We isolated visual pathways by presenting flickering (on-off) grating stimuli modulated along different dimensions in MacLeod Boynton colour space: (L+M+S, L−M and S-cone isolating) and recorded steady-state visually evoked potentials at three different modulation frequencies (5, 12 and 16 Hz). We then trained a support vector machine to classify chromaticity using the pattern of either complex or magnitude-only frequency domain responses and measured classification accuracy across the frequency domain (1–100 Hz) in bins of 1 Hz for each participant.
Classification based on incoherently-averaged signal power was significantly above chance across the much of the frequency range. Further analysis revealed this is due to chromatically-dependent changes in broadband power. Classification accuracies based on complex phase were significant mainly at harmonics of input frequencies with significant classification accuracy even at very high harmonics (>60Hz). In addition, we found clusters of statistically significant intermodulation terms that hint at a phase-sensitive interaction between endogenous activity and chromatic inputs in the early visual cortex.
This PDF is available to Subscribers Only